Transforming Sequence Tagging Into A Seq2Seq Task

TL;DR

This paper systematically studies how to best format sequence tagging tasks as Seq2Seq problems, introducing a new format that improves accuracy, robustness, and multilingual transfer, supported by extensive experiments.

Contribution

The paper provides a rigorous analysis of input-output formatting for Seq2Seq sequence tagging, proposing a new format that outperforms existing ones across various benchmarks.

Findings

The new format simplifies the task and improves accuracy.

It enhances multilingual zero-shot transfer and joint training.

It reduces hallucination and increases robustness.

Abstract

Pretrained, large, generative language models (LMs) have had great success in a wide range of sequence tagging and structured prediction tasks. Casting a sequence tagging task as a Seq2Seq one requires deciding the formats of the input and output sequences. However, we lack a principled understanding of the trade-offs associated with these formats (such as the effect on model accuracy, sequence length, multilingual generalization, hallucination). In this paper, we rigorously study different formats one could use for casting input text sentences and their output labels into the input and target (i.e., output) of a Seq2Seq model. Along the way, we introduce a new format, which we show to to be both simpler and more effective. Additionally the new format demonstrates significant gains in the multilingual settings -- both zero-shot transfer learning and joint training. Lastly, we find that the new format is more robust and almost completely devoid of hallucination -- an issue we find common in existing formats. With well over a 1000 experiments studying 14 different formats, over 7 diverse public benchmarks -- including 3 multilingual datasets spanning 7 languages -- we believe our findings provide a strong empirical basis in understanding how we should tackle sequence tagging tasks.